KyutaiDactory
Dactory language classifiers
Dactory provides trained fastText classifiers used to filter Common Crawl text. The release covers the 24 official EU languages and assigns content labels that help select data for Helium training.
Research release
Most labels use translated training data
The classifier training data is translated with MADLAD for multilingual use. Wikipedia and random-web examples can instead come from each language. These are text classification models, not generators or chat assistants.dactory-models: Kyutai model card
Specifications & access
| Detail | Value and conditions |
|---|---|
| Status | Research release |
| Access | Downloadable |
| Model identifiers | kyutai/dactory-models |
| Weight license | CC BY-SA 4.0.dactory-models: Kyutai model card |
| Labels | rand, wiki, books, science, stem, hum, pop and life.dactory-models: Kyutai model card |
Related models
- Helium 1 2B
Kyutai ยท Helium
A 2B base language model for the 24 official EU languages, built by combining selected trained variants.
Launch date not confirmedResearch releaseDownloadable